RAID and Game Loading

[Sjaak]

I was really surprised by the results and conclusions posted in this article. I am using raid0 for about a week now, and i have had only (extremely) positive experiences. All loading times decreased, sometimes by more then 50%

In the frontpage article, times of loading have been measured. I was astonished, especially about this:

Far Cry - I timed the load up of the first level after the video

Single 36Gb Raptor: 1.30.42 (one and a half minute)
raid w/ 64k Stripe: 1.23.70 ( one minute 23,7 seconds)


Couldnt believe this - so i timed my own loading:

31 seconds until able to play (move character) Also, times in other games are much faster then said in the article...i think something is wrong somewhere.

To me, RAID0 really brought big advantages. Load times in C&C Generals almost disappeared (less then 1 second), and UT 2k4 is only a matter of a few seconds. I think this author (no personal offence!!!) published non-accurate results and thereby gives RAID a bad name. What times do other raid users get??

 

[1-Man-Army]

hmm even tho i dont have raid any more, or the times recorded, i just wanted to say, that when i used to have raid, my internet conection was also faster, and much more stable, instead off fluctuation like crazy with a single hdd.....but that was when i had 2x 13gb hdds soo i dont no about now...besides that i did feel faster load times with raid too dont have the times tho sorry

 

[Sonny]

The testing was done with the install imaged with Norton Ghost & restored with it. Norton Ghost doesnt officially support RAID0 even though it might work in some cases. Are we seeing flaws in the restore image causing the results to be skewed? To have a reliable test Norton Ghost should not have been used.

 

[Vio1]

My load times were not effected by any error with Norton Ghost... otherwise there would have been a drastic difference between my first 2 tests (16k and 32k) because 32k would have been the first time I used the image made by norton ghost.
As to why Far Cry takes so long to load, its because I started timing the moment the bar appeared to the moment the beginning of the video appears (and you see your character lying on the bottom of a hole).... plus, the amount of time is not important.. its how the times change or do not change between tests .... thats what is important.

 

[Sjaak]

well, with my old disk setup (single 7200 pata), that loading time was about a minute, if i remember right. Big advantage, to say so.

 

[Sonny]

My load times were not effected by any error with Norton Ghost... otherwise there would have been a drastic difference between my first 2 tests (16k and 32k) because 32k would have been the first time I used the image made by norton ghost.
As to why Far Cry takes so long to load, its because I started timing the moment the bar appeared to the moment the beginning of the video appears (and you see your character lying on the bottom of a hole).... plus, the amount of time is not important.. its how the times change or do not change between tests .... thats what is important.

What's important is that the test procedure conducted be unquestionable so that the results have a basis in reality. I don’t trust the differences that you claim to be important since the way they were obtained are unreliable.

Testing is time consuming; the process needs to be developed properly without any regard for length of procedure or what’s easy for the tester.

 

[Vio1]

How is Norton Ghost unreliable? The testing was already very time consuming... and wouldnt installing a fresh copy of windows and fresh copy of every game each time I do the test be more likely to have inconsistencies?

I dont follow your logic how using an image of your HDD and then copying it makes it unreliable.

 

[Sjaak]

if you copy an entire HDD the FAT tables etc. might get screwed up - im not known in that area of pc'ing, but it seems very likely to me that alot can go wrong there.

 

[Vio1]

if something did infact go wrong... wouldnt it be more evident? Such as with odd results??? What results I came up with parallel the results that others have (all the reviews and hoop la circulating the web).

 

[1-Man-Army]

hmmm seeing as the question off norton ghost is reliable is brought up.. i wonder, when u copy the image put it on another hdd if it affects the fragmentations....or the order off the files... u could have also used fraps to get a almost perfect timming by recoding it kinda like a movie... but i think that might slow things down... not sure

 

[Sonny]

Situation:
You want to know whether Ghost is compatible with RAID.

Solution:
Ghost is not compatible with computers that use RAID. That is, Symantec Ghost 7.5 and earlier, and Ghost 2003 and earlier, do not support RAID controllers on computers that are being imaged. In addition:

1. Ghost does not work with software level RAID.

2. Although Ghost might work in some limited circumstances with hardware level RAID, Symantec does not support using Ghost for cloning hardware level RAID drives.

3. Ghost does support cloning drives that have had RAID removed, such as after you break a mirrored set.

That's from Symantec.

 

[DocClock aka MadClocker]

I would think that the best test scenario would be to have all nessessary drivers and such, and do a fresh install for each test. If you have done one install and turn around and setup raid and install again as long as you don't change anything, the only thing different will be the extra hdd space reported durring setup...and with that in mind, there might have been a difference in cluster size... which might make a small difference also

 

[Sjaak]

that would be best, yeah, but that would take an unholy amount of tme. I'm deciding based upon mine and other peoples direct experience, and i say its worth it

 

[theELVISCERATOR]

I have retail ghost 2003 and it wouldnt work for my raid AT ALL

 

[shadowdr]

I have to agree with Sonny.I have used Ghost every since i started using raid0.It is a good backup strategy as long as there is at least one more like CD or dvd backup.There have been times when the Ghosted image would not work and I had to use the dvd to get my partition back.It may be the versions that I have used or the way DOS treats long file names but when I do rewrite a partition with Ghost I also must fix some errors that come with it.First I run chkdsk from the recovery console.It always finds errors within the files and often runs pages and pages of re-association before it is done.After rebooting, I always run Norton's Win Doctor.It finds anywhere from a few to a few hundred errors and I let it repair them without even checking the details so I cannot be sure what exactly they are but there are always some.

I still consider it my best backup option as opposed to losing my OS or program partitions,but it is not perfect.

 

[Imperial]

I've never really cared to much about RAID. Who cares about seven seconds?

 

[jlin453]

I've never really cared to much about RAID. Who cares about seven seconds?

For users who do video editting and such, raid0 allows faster transfer speeds. Not everyone using raid0 is doing it for "seven seconds" of game loadup.

 

[Admiral Falcon]

Yes, for sustained data transfer, RAID-0 really shines. I don't have a setup myself, but I do know that it theoretically doubles your sustained transfer rate. For things like transferring and editing video, this is a huge deal. DV camcorder video runs about 14GB per hour. That's a lot of data to be moving around, and the faster you can do it, the better.

 

[Mr.Guvernment]

i love SCSI raid 0 with 15k cheetah drives - i can extract about 30 + zip files @ 200mb each with top speed and they are all done in like 20 mins!

 

[Sjaak]

I've never really cared to much about RAID. Who cares about seven seconds?

if it saves me 7 seconds a 50 times a day, thats almost 150 hours in a year